The Bronze Age and the Celtic World

FOR the last fifteen hundred years the Celtic tongues have been spoken only in the extreme north-west of Europe, in parts of Ireland, the west of Scotland, and the Isle of Man, in Wales, Cornwall and Brittany, and for some little time these languages have ceased to be spoken in Cornwall and the Isle of Man.

But we have ample evidence that these tongues had once a wider range, and were pushed westward in the first instance by the spread of Roman culture and the Latin language as the empire increased its bounds, and still more by the Teutonic tribes who invaded the western half of that empire and brought about its fall.

If however we examine the evidence which has come down to us from the first century before the Christian era, especially such material as has been furnished by C?sar and Strabo, we shall find that languages of the Celtic type were spoken at that time throughout all Europe west of the Rhine and north of the Pyrenees and the southern slopes of the Alps. We shall note also that these tongues were spoken in many parts of Spain and in North Italy, though in both these areas they were of relatively late introduction.

Again there is another area in which Celtic speech was in use at that time, or had been shortly before. This is the mountain or Alpine zone of Central Europe, as far east, at least, as a line drawn from Cracow to Agram. It is possible, too, that such tongues may have been spoken at one time still further east.

The problem before us is to inquire first in what region the Celtic tongues originated, then how and when they spread to the areas in which we find them two thousand years ago. To do this we shall have to review the condition of these areas both from the standpoint of prehistoric arch?ology and physical anthropology, to see whether the evidence derived from these sciences, taken together with that drawn from comparative philology and the study of place-names, can help us to reach a solution.

But the problem is further complicated by the fact that the Celtic languages fall into two groups. In the one occurs the sound qu, which has in later days become a hard c, while in the other this sound has become labialised and converted into a p or b. It has been thought by some that the qu peoples, spoken of usually as Goidels or Gaels, arrived first from the common Celtic home, and that the p peoples, called Brythons or Cymri, came later from the same centre; this view is, however, strenuously denied by others. We have, therefore, to determine if we can, not only whence and when the Celtic languages arrived in the west, but whether they came in one, two or more waves.

FIG. 1.-CELTIC LANDS AND THE CELTIC CRADLE.

Lastly, we find that the Celtic tongues, as spoken to-day, contain elements of grammar and syntax, and not a few words too, which divide them off sharply from those groups of languages to which they are in other respects akin. Also it is believed by some that non-Celtic languages, such as Pictish, survived in this region until relatively late times, while it is well-known that a primitive non-Celtic tongue, the Basque, is still spoken in the fastnesses of the Pyrenees. It is important, therefore, if we are to have before us all the factors which enter into the problem, that we should inquire what people were here before the first Celts arrived, and that we should make ourselves to some extent familiar with all the different races and cultures which preceded the Celtic invaders.

If we pass across England and Wales from east to west, and the same is almost as true if we cross Scotland, we find, first of all that the population is mainly tall and fair, while as we proceed we come across elements which are darker and shorter, until in Wales and the West Highlands we find the majority of the people are small brunettes of slender build. This dark type is also to be met with in Ireland, especially in the west, the part of that island in which the Erse language has best survived.

It is because the Celtic tongues, whether qu or p, are spoken chiefly by people of this small brunette type, that it is frequently called the Celtic race, and yet all the evidence of ancient authorities goes to show that 2,000 or 2,500 years ago the Celts were looked upon as a tall, fair people.[1] Here is another difficulty which must be taken into consideration as we make our inquiries, for no solution can be considered sound which cannot, without straining the evidence, answer all these questions.

As we have seen the main areas which were Celtic-speaking in the time of C?sar were the British Isles and Gaul, west of the Rhine; these I shall term Celtic lands, leaving out Spain and Cis-alpine Gaul as areas into which the Celtic invasion arrived at a relatively late date. Now, besides these Celtic lands Celtic tongues were spoken in the Alpine zone, and perhaps at one time still further east. It is from this area that the Celtic languages have been thought by some to have entered the lands of the west. They cannot have been introduced from Spain or Italy, into which they were late entrants, but it has been suggested by some writers that they arrived from the north-east, from the Baltic region. It is true that there is some slight evidence that Celtic place-names have existed in this area, but the balance of evidence, as I shall hope to show, seems to prove that Celtic people arrived there relatively late and not in large numbers, and that they were never the dominant people of that region. There remains only the Alpine zone and the lands to the east of it. This area, from the Jura to the Iron Gates, from the northern slopes of the Carpathians to the southern foot-hills of the Alps, I shall term the Celtic cradle, and I trust that the evidence which I shall produce will convince my readers that I am correct in so doing.

* * *

OF the earliest inhabitants of Celtic lands we know little or nothing. We have, it is true, a number of tools made of flaked flint, but they tell us little of the men who fashioned them. In spite of the recent admissions by the eminent French arch?ologists who have examined the new discoveries at Foxhall,[2] there is still no little difference of opinion as to the human workmanship[3] of rostro-carinates, eoliths and such like early attempts, and no human remains have come to light which can be attributed with any probability to this horizon.

When we come to what is usually termed the lower pal?olithic period we are on surer ground, for no one now denies the origin of implements of the Chelles and St. Acheul types. But the only skeletal remains which can with certainty be attributed to this period are the human jaw from the Mauer sand-pit near Heidelberg,[4] and the famous Piltdown skull.[5] Few people now believe that the Galley Hill skeleton dates from so remote a time,[6] while the discoverer himself has disclaimed so early an origin for the Ipswich man.[7]

To attempt to reconstruct a human type from a mandible alone would be indeed to carry far the principle of ex pede Herculem, and as yet there is little agreement among anthropologists as to the exact date, or for that matter the exact reconstruction, of the Piltdown skull,[8] though the ingenious hypothesis that a unique human cranium without a jaw, was found in close association with a unique troglodyte mandible has now, I understand, definitely been abandoned.[9]

Thus little or nothing is known of the first inhabitants of Celtic lands, beyond their tools, but when we come to the middle pal?olithic period the case is different. While some difference of opinion still exists, the view advanced by Obermaier[10] and others seems to be gaining ground, that in Celtic lands the industry of Le Moustier first appeared as the climate was becoming colder on the approach of the last or Würm glaciation, though it is thought by some that it had flourished in an earlier and warmer time in the regions lying to the east.[11] This industry is believed by most authorities to have survived the first Würm maximum and to have lasted through the temporary amelioration of the Laufen retreat. Whether it survived, too, the second maximum, and lasted until the climate definitely improved is more doubtful, but many arch?ologists of great repute believe that it did so,[12] and unless this was the case it will be difficult to explain certain features of the Audi flints.[13]

Though there is as yet no general agreement as to the duration of the Mousterian industry, it is different when we come to consider the type of man who was responsible for this work. Everyone is agreed that the authors of this culture were of the type known as Neanderthal man, for several skeletons of this type, or parts of them, have been found associated with flint implements of Le Moustier design, and none have as yet turned up under conditions which make this correlation impossible.[14]

A considerable number of skulls and skeletons, about two dozen in all, of Neanderthal man have been found, the great majority in Celtic lands; but, though there is a general resemblance between all the members of the series, sufficiently strong to mark them off from the Piltdown skull on the one hand and from modern men on the other, the type is in many respects very variable. There are vast differences observable between the skull from Chapelle-aux-Saints,[15] the highest form yet discovered, and that of the Gibraltar man,[16] or rather woman, which is the most primitive yet found in Europe. As far as one can judge from the descriptions which have appeared as I write, the skull recently found at Broken Hill in Rhodesia differs from that of Gibraltar hardly if at all more than the Gibraltar skull differs from that found at Chapelle-aux-Saints. In the latter case there are several intermediate forms, in the former such may yet turn up, for Africa has, as yet, produced but one other skull of this type, that found not long ago near Constantine in Algeria, no description of which has, I believe, yet been published.

Skulls of this type have been so frequently described,[17] individually and collectively, that it is unnecessary to give another detailed account. It will be sufficient to say that they are large and massive, the vault is low, and they are specially distinguished by having over the eye sockets a heavy and continuous projecting ridge, known as a torus, which is one of the distinguishing features of the large anthropoid apes. Another point of importance is that the head was so attached to the body that it could not have been held absolutely erect, and must have produced a slouching gait, though the degree of this slope varied considerably in different specimens, and in the case of the Rhodesian skull was quite halfway between the slope of the Gibraltar skull and that of the gorilla.[18]

But it is unnecessary for our purpose to pursue this question further, for with the arrival of modern man, after the last glaciation was past, Neanderthal man disappeared. That the two races met, though not necessarily in this continent, seems clear from the fact that at Audi, near Les Eyzies, in the Dordogne, we find a culture, which in some respects resembles that of Le Moustier, and in others the succeeding culture of Aurignac.[19] That these two races interbred is unlikely, for Neanderthal man must have appeared an unsightly beast to his modern successor. In any case, if mating did take place, the union must have been sterile, for, in spite of much that has been written to the contrary,[20], there is no clear evidence of the survival of any distinctive Neanderthal traits in the men of later days.[21]

The second maximum of the Würm glaciation seems to have culminated about 15,000 B.C.,[22] and about that time, or conceivably earlier, modern man first arrived in North Africa, if we may judge by the appearance of a fresh type of flint industry, known usually as Capsian.[23] Whence he came is uncertain. It has been suggested that he may have reached the north from tropical Africa,[24] but no evidence has been adduced in support of this hypothesis. It seems more likely that he came from Asia, probably by means of the Sinaitic peninsula, or possibly across the Straits of Bab-el-Mandeb. This much is certain; about this time the Capsian culture is found extending along the north of the continent, from Egypt as far west at any rate as Algeria, and perhaps beyond, though at no point but one is it found far from the Mediterranean coast.[25] The one exception is in Egypt, where implements of this type have been found as far south as Luxor,[26] so that we may be satisfied that modern man in his earliest movements passed up the Nile valley at least as far as the First Cataract. It would seem probable that in Egypt the invaders came into touch with their Neanderthal predecessors, who retreated before them up the Nile valley towards Luxor, where Dr. Seligman has found implements of Le Moustier type more developed than any discovered elsewhere[27]; it is possible that some retreated further south and may even have reached Rhodesia.

Other of these Neanderthal refugees seem to have gone westward, and perhaps passed up the Italian land-bridge to western Europe; if so it was probably these, who had come into contact with the Capsian culture of North Africa, who were responsible for the Audi industry. They were followed before long by the invaders, and in Celtic lands at least were soon exterminated, though it is just possible that they survived to a later date further east.[28]

The culture of the newcomers is known as that of Aurignac, and seems to have started in Europe about 12,500 B.C. A great many skeletons of this period have been discovered and described, and though all of these show us men very like those of the present day, there is a considerable range of variation among them.[29] The skulls of the upper pal?olithic periods, apart from the Chancelade skull[30] to be discussed later, may be divided into three marked groups, though it is well to remember that there is no strict uniformity among all the members of each group. All the skulls of this period, however, are long, for the broad-headed type, so prevalent in Central Europe to-day, did not arrive until the closing phase.

Of the first of these three groups we have only two examples, the mother and son from the Grotte des Enfants, near Mentone.[31] But as these are the earliest in date, and differ in some respects very markedly from the remainder, they have been distinguished by the name of the Grimaldi race, after the owner of the cave, the Prince of Monaco.

This type was small, being less than 5 ft. 3 in. in height, the skulls were of the long variety, having length-breadth indices of 68.5 and 69.2, and the jaws and teeth project, so that they exhibit a character known as prognathism. This latter character has caused the race to be termed negroid, and unjustifiable deductions have been drawn from this term. It has been shown, however, that there is no reason for supposing any affinity between this type and the negro race of tropical Africa.[32] Both of these skeletons were found in a contracted position, and that of the boy was covered with red ochre.[33]

Our second group is the Cromagnon, and is based largely on the skeletons found in the cave of Cromagnon, near Les Eyzies. By many anthropologists this term is used to cover all the skeletons from this period except those of the Grimaldi type, but more recently it has been shown that all these remains cannot conveniently be placed in one group, for the distinguishing characters are but faintly visible in some and totally absent from a large number.[34] The term is now becoming used in a more restricted sense.

The Cromagnon type is tall. The men were often 5 ft. 10 in. or 5 ft. 11 in. high, though the women were frequently much shorter. Their heads were large, larger than the average in Europe to-day, but not very high; they were long as compared with their breadth, having a cranial index of about 74; their noses were narrow, but their faces were short and relatively broad. This combination of a long head and a short face is unusual, and is called disharmonic, and this disharmony is one of the most striking characteristics of Cromagnon man.[35]

It is often thought that this disharmonic trait, the long head and the short face, is evidence of the mixed ancestry of the race which exhibits it,[36] and if this were the case we might expect Cromagnon man to be the result of a crossing of two other races. There is no other evidence to indicate that this was the case, and if such crossing had occurred, it seems likely that it took place before the Cromagnon type reached Europe.

It seems probable that it is to the men of the Cromagnon type that we must attribute the beginnings of that art, which reached its finest development in a later age, and has provided the most conspicuous as well as the most pleasing feature of the upper pal?olithic culture.[37]

Lastly we have the type represented by Brünn I., Brüx, Lautsch, Combe Capelle, Barma Grande (one of the skulls from B.G. now in the Musée de Menton, but not the skulls generally known as B.G. 1 and 2), the woman from the upper layer in the Grotte des Enfants, the Calotte du gravier de fond at Grenelle, the Denise fragments, as well as by one or two skulls of the transition period from pal?olithic to neolithic found at Ofnet (No. 21, i.) and a few of those belonging to the same period found at Mugem. The type is usually high-headed as well as narrow-headed, and tends to have the orbits horizontally lengthened, the glabella and supraciliaries strong, the fore-head retreating, the nose broad and the upper jaw projecting (alveolar prognathism). The cephalic index is usually between 68 and 72; the stature is moderate or low.[38]

Thus we find during the period of Aurignac three groups of long-headed men, the Grimaldi, Cromagnon and Combe Capelle, and, especially on the Riviera, in the Barma Grande cave and the Grotte des Enfants, skulls which show various apparent combinations of these types, while at Solutré and Laugerie Basse we find the last type showing modifications to some extent towards the characteristics of modern men. These types and intermixed types occupied west and central Europe, so far as it was habitable during the later pal?olithic periods, and the combinations of Combe Capelle and Cromagnon characters in the skulls of Obercassel (Magdalenian period) is noteworthy. The earliest in point of time is the Grimaldi, which has been found only near Mentone, and there are reasons for believing that its distribution lay around the western Mediterranean, then an inland sea. This view is supported by the fact that marked alveolar prognathism has been noted among the natives of Algeria and Morocco, and I am told that it is not uncommonly met with in Spain; it is also very marked in Portugal, though here it has been attributed to a different cause. It is, however, of old standing in that country, as it has been noted among the skulls from Mugem,[39] which are believed to date from the close of the pal?olithic age. A similar feature has been noted in some of the skulls from the Algerian dolmens.[40]

To the Cromagnon type, pure, it is difficult to ascribe any other skulls besides those from Cromagnon, and those from Lafaye Bruniquel, but some of the Cromagnon characters are well shown in some Barma Grande skulls. The type is said to survive in the Dordogne and perhaps near the western Pyrenees in North Spain at the present day.[41] The Combe Capelle or Brünn type, is seen to have occurred on the whole more to the north and east, and seems rather to focus in Central Europe and the southern part of the North German plain. It was probably the latest to arrive on the scene, for it is associated only with remains of late Aurignac type, and has been more frequently found in the succeeding Solutré period.

Thus we see that by the close of the period of Aurignac, about 11,000 B.C., we have three groups of long-headed men in Celtic lands, and that, though they overlap, they are tending to obtain for themselves definite areas of distribution.

During the closing years of the Aurignacian period the climate had been getting milder and perhaps drier, and steppe conditions prevailed over much of France and still more further east. Herds of horses arrived and were hunted for food and the saiga, a kind of antelope, was found as far south as the Dordogne, if not beyond, during the succeeding Solutrean period. These Steppe conditions are more characteristic of the latter period,[42] when France was invaded by a new people, not given, as far as we know, to artistic efforts, but who were able to fashion very skilfully made weapons of flint to aid them in chasing the beasts of the steppe.[43] The fact that skulls of our third group the Combe Capelle, are more common during this period and have only been found during the later phases of the previous age, when, as we have seen, steppe conditions were already approaching, leads us to suspect that it is to this type of man that we must attribute the invasion of Celtic lands which took place at this time. The Cromagnon men seem to have retreated to the south-west and to have taken refuge in the fastnesses of the Pyrenees,[44] while the invading hunters dominated the southern part, at least, of the Celtic lands.

But towards 9,500 B.C. the climate began again to deteriorate, and the steppe conditions passed gradually to those of tundra. The steppe animals retreated to the east, towards South Russia and Turkestan, and most of the men of Solutré, who hunted them for food, seem to have followed in their wake. It seems doubtful whether the Solutrean invasion reached Britain, though implements of this type are said to have been found here,[45] and Proto-Solutrean stations are reported as occurring in England.[46] It has been claimed recently that this type reached the south of Sweden,[47] but this view is not generally accepted in that country.[48]

On the departure of the Solutrean invaders the remnant of the aborigines, who had fled to the mountains in the south-west, and there developed their art to a much greater pitch of perfection, now returned to France, and once again, as the men of La Madeleine, became the dominant race in Celtic lands. It seems possible that some of their comrades had fled north to Britain on the arrival of the men of Solutré, and had survived there throughout this period, for, though no industry has been found in the British Isles which can accurately be described as that of La Madeleine,[49] in the strict French meaning of that term, we do find traces of the culture of Aurignac, persisting perhaps until still later times.

It must not be thought, however, that the Combe Capelle race never reached these isles. Whether the culture of Solutré did so or not seems uncertain, but some of the skeletons which have been found here have been classed with the Combe Capelle group.[50] But, as we have seen, this race was present in France, at any rate in some parts of that country, for some little time before the arrival of the men with the culture of Solutré.

The colder climate of the Magdalenian period has been shown to coincide with the Bühl advance of the Alpine glaciers,[51] which reached its maximum about 7,500 to 7,000 B.C. After that the climate slowly improved, though the precipitation increased, and forests sprang up on the hitherto open lands. As the tundra conditions in Celtic lands gave way to forest, the reindeer migrated to the north and north-east, while their place was taken by the red deer. As the forests developed it became increasingly difficult for men to traverse great distances or to intermingle as freely as they had done before. There was a tendency for separate groups to develop in different regions; so that, when we arrive at the next period, the Azilian, we find very different types of people in various parts of Europe.

Even before the close of the Magdalenian period a fresh type had arrived, apparently from the north, if we may judge from the skeleton found at Chancelade in the Dordogne. This skeleton bears a close resemblance to those of the modern Eskimos,[52] and since the latter have retained a type of art reminiscent of certain phases of Magdalenian culture,[53] we may suspect that Chancelade men, following the departing reindeer, passed north-eastward to the tundra of Siberia.

It was between 7,000 and 6,500 B.C. that a fresh wave of Capsian people from North Africa began to invade Spain,[54] into which peninsula they introduced what is known as East Spanish Art.[55] By degrees they pressed the Magdalenian Cromagnons to the Pyrenees, where their culture declined to that which we know as Azilian.[56] The invaders passed on through Celtic lands, bringing with them a new culture, known as Tardenoisian,[57] and seem to have reached the British Isles before 5000 B.C.

These people seem to have been another variety of the same long-headed race, which had developed into a distinct type in North Africa, and had there, perhaps, mingled to a greater or lesser degree with the descendants of the Grimaldi men, whom we met with at the beginning of the period of Aurignac. If we may judge by those who seem to be their descendants, they were of rather short, slight build, with long narrow heads, brown skin, dark hair and eyes, the type which to-day is known as the Mediterranean race.[58] It is possible that the Grimaldi elements in their composition, and which are sometimes found comparatively pure, may account for that small dark type, often showing marked alveolar prognathism, which has been found in certain out of the way regions, such as Apulia and Sardinia, and which are known to some anthropologists as Iapygian,[59] and have been termed Ethiopic by Ruggeri.[60]

This new population seems to have been peaceably inclined and made no attempt to exterminate its predecessors, but settled down in the lower lands and by the sea shore, while the Cromagnon men remained in the mountain zones of the Pyrenees and the Dordogne, and the Combe Capelle type survived in Central Europe and among the hills of Wales. It seems almost certain that the newcomers were still hunters, quite ignorant of agriculture and the domestication of animals; as some of their settlements have been found by the sea shore and on the banks of streams, it seems likely that they lived to a considerable extent on fish and molluscs.

It would appear, then, that the type which we know as the Mediterranean race, and which has given to Wales, Scotland and Ireland the majority of their small brunette inhabitants, is made up of the descendants of all the types of long-headed men-except the Chancelade variety-which we meet with in the Celtic lands of western Europe during the upper pal?olithic period. That the Combe Capelle type survives on the moorlands of Plynlimmon has been shown by Fleure: examples of an africanoid type with alveolar prognathism are not uncommon in Wales and in the poorer quarters of our big cities, and the Cromagnon type only seems to be missing or at any rate relatively scarce. The main element, however, which has gone to make up the Mediterranean race as we now know it, seems to be that which entered Europe through Spain, with Capsian culture, during the closing years of the Magdalenian period.

These people have left in the west, not only considerable vestiges of their blood, but no small amount of their language, or to state the matter more accurately the language of these people has left a marked effect upon the tongues which succeeded it in the west. More than twenty years ago Mr., now Sir John Morris Jones[61] pointed out that "the syntax of Welsh and Irish differs in some important respects from that of the languages belonging to the other branches of the Aryan family," and suggested that these points, in which too the neo-celtic tongues differed from ancient Gaulish, were due to the influence of a language which had been spoken in these lands before the introduction of the Celtic tongue. He pointed out that many of these peculiarities, which occur also sometimes when the English tongue is spoken by Irishmen, were similar to the syntactical arrangements in force in the language of ancient Egypt and among the Berber dialects spoken by the natives of Algeria, the Kabyles, Shawiya and Tuaregs.

Now the Egyptians and other peoples of North Africa are considered by all anthropologists as typical members of the Mediterranean race, though the inhabitants of the western part seem, as we have seen, to have incorporated no small amount of Grimaldi blood; it would seem then that we may accept the suggestion of Sir John Morris Jones that the syntax of Welsh and Irish is a legacy from the language spoken by these Mediterranean invaders, who reached Spain about 7000 B.C. and formed the bulk of the population of the British Isles about 5000 B.C.

So far we have been dealing with the early inhabitants of the Celtic lands of the west, but a word must be said of some fresh arrivals into the Celtic cradle in Central Europe. It was during the Azilian period, about 6000 B.C., that a new race appeared in Central Europe, coming from the east. Of their earlier abode we know nothing positively, but there are reasons for inferring that their line of approach was by the Kopet Dagh and the Armenian highlands, and that they came ultimately from the slopes of the Hindu Kush and the western side of the Himalayan massif. This race, which is called the race of Ofnet, from the skulls found in the caves of Ofnet, in Bavaria, had a broad head, the outline of which as viewed from above consisted of two segments of circles, the one forming the back of the head, the other the front. The brow-ridges are slight, the nose short and straight, the eye-sockets low and almost rectangular, the cheek-bones not very prominent and the chin weak and undeveloped.[62] This race seems to have met and mated with the remnants of the Combe Capelle race in the Upper Danube basin, and the progeny of this union seems to have been a type with a pear-shaped head as seen from above, with a rounded back, indistinguishable from the type found later in the Swiss lake-dwellings and in the mountains of Central Europe at the present day, and which is known as the Alpine race.[63]

The Ofnet race seems to have spread westward into the Celtic lands, either at this time or perhaps later, though probably in small numbers, for a skull found at Grenelle, near Paris, under what are believed to be neolithic surroundings, belongs to this type.[64] Other broad-headed skulls of this or the Alpine type, dating from about 5000 B.C., or a little earlier, have been found at Mugem on the banks of the Tagus,[65] while others of this type of about the same date have been found in the caves of Furfooz in Belgium.[66]

Whether any of this broad-headed Asiatic strain reached the British Isles at so early a date is uncertain. No skulls of this type and date have been discovered, but broad-headed types occur sporadically in Wales, Ireland and the western islands of Scotland, which may conceivably represent descendants of early Ofnet or Alpine immigrants.

Somewhat later, before 4000 B.C., fresh waves of broad-headed immigrants seem to have arrived in Central Europe from the Armenian highlands or the Anatolian plateau, bringing with them the knowledge of grain, cultivated fruits and domestic animals, and the custom of erecting pile-dwellings in marshes or lakes, and of grinding and polishing axes of flint or other hard stone.[67] Such knowledge seems to have reached even the west of Switzerland by 4000 B.C. and to have spread later throughout the massif central of France, which was already peopled by men of the Alpine type.

But the art of polishing hard stone spread further than the people who were responsible for its introduction, and during the next few centuries this art had become well known throughout the Celtic lands of the west; the need for more efficient tools to fight the encroaching woodland must have encouraged this art. How far the elements of agriculture had travelled with the art of grinding axes seems uncertain, for few, if any, unquestionable neolithic dwelling sites of this time within this area have been found or thoroughly explored. The scanty evidence at our disposal seems to show, however, that the people of the west were possessed of some domesticated animals, so that the inhabitants of Celtic lands had passed from a purely hunting stage before 3,000 B.C.

There is one other culture, introduced into Europe perhaps by another race, which I must not omit to mention, as it may have provided another element, albeit a small one, in the early population of Celtic lands. At Mullerup, in the peat moss of Maglemose, in the west of the island of Zealand, there was found in 1900 an important dwelling site with a very distinct culture, including harpoons and other implements of horn and bone, which is known to Scandinavian arch?ologists as the Mullerup, but to English-speaking students as the Maglemose culture.[68] More recently, in 1917, another settlement, exhibiting what appears to be the same culture, was discovered at Sv?rdborg, in the south of the same island.[69]

No skulls or skeletons have been found associated with this culture, and there has been much speculation as to the race which was responsible for it. Owing to the presence of harpoons it was first assumed that this culture was a direct derivation from the Azilian and Magdalenian, though it has been pointed out that the Maglemose harpoons are very different in form from the Azilian, and resemble more nearly some found in eastern Russia.[70] Still the majority of authorities treat this culture as of Azilian origin. Others, relying largely on the resemblances of certain elements of culture to those found at some very late Aurignacian sites in South Poland, believe the people and the culture to have arrived from that region.[71] Recently I have suggested another explanation.[72] Noticing the resemblance between the Magiemose culture and a slightly later civilisation known as East Scandinavian or Arctic, which has been found at several sites associated with skulls of Mongoloid type, I have suggested that in the Maglemose people we may perhaps see the first arrival in Europe of that Mongoloid race, which now peoples a large part of the north-east of the continent. My suggestion has not been well received in Scandinavian circles, and M. Nordmann has submitted it to very searching though courteous criticism.[73] While duly appreciating the value of all the evidence he has cited, I am still of opinion that my view, though far from proved, meets the existing evidence as well as, if not better than, its rivals.

The importance of the Maglemose problem for our purpose lies in the fact that certain sites in the British Isles have produced an industry which has been claimed, and perhaps rightly, to be akin to that of Mullerup and Sv?rdborg. Certain discoveries in the caves at Oban and on a raised beach on the island of Oronsay, are claimed to be of this or of Azilian culture,[74] while other finds at Holderness are said to resemble more closely still the Maglemose culture.[75] More recently still Mr. O. G. S. Crawford has suggested that certain implements, which he and I discovered last year at an early occupation site on the Newbury Sewage outfall works at Thatcham, Berks., bear close resemblances to some found at Sv?rdborg.[76]

It is too soon yet to appraise the value of these resemblances. Some of these sites, notably those at Oronsay and Thatcham, appear on some grounds to be somewhat later than the settlements at Mullerup and Sv?rdborg. This does not, of course, disprove their cultural connection. It is unwise, at present, to draw any positive conclusions from such evidence, but we may note that it is possible that during late Azilian times, or perhaps later still, fresh elements entered the British Isles from the Baltic region, and that it is at least possible that these elements may have been of the Mongoloid race.

People showing slight Mongoloid traits may be found sporadically throughout Wales, though, as far as I can ascertain, this type has not been noted as prevalent in any particular areas; how far it may be noted in the west of Scotland or in Ireland I am uncertain. But we cannot be sure that the introduction of this Mongoloid strain dates from so early a time, as it is quite possible that the type may have been introduced much later by the Vikings, who may perhaps sometimes have carried Finns with them in their forays.

Though after the close of Azilian times the culture of Celtic lands changed more than once and in more respects than one, we have at present no reason for suspecting the introduction of fresh racial elements before the beginning of the Bronze Age. The origin of Campignian culture, which seems to have flourished over the northern part of Celtic lands, in one form or another, from about 5000 to 3500 B.C., is still a matter of dispute, but it is doubtful whether the solution of the problem is likely to introduce a fresh element into the population of the Celtic lands.

The vast mass of the population of this area about 3000 B.C. were the descendants of the long-headed populations of Europe and North Africa in the upper pal?olithic period. In some parts of the south the Cromagnon type may have persisted, in a pure or mixed form, as did the Combe Capelle type further north, while a modified form of the Grimaldi type was found from Portugal to Wales, especially in fishing villages. The prevailing type seems to have been that which came latest from Africa, and which most truly deserves the name of the Mediterranean race, though it may be well to realise that this term, as commonly used, seems to include all the varieties before mentioned, as well as a modified mixture of all these long-headed types.

In Central France, and to a less extent elsewhere, the Alpine type had penetrated, though it is doubtful whether it had, as yet, reached the British Isles. And we must realise that it is just possible that some Mongoloid peoples, from the Baltic and ultimately from Siberia, may have made a few settlements in this country, though their numbers are not likely to have been great.[77]

Such then, as far as our evidence extends at present, seems to have been the population of Celtic lands in the true neolithic age, when people lived in small, self-contained communities, and outside commodities were rarely met with, and then only bartered from tribe to tribe. As we shall see, the next thousand years or so were to introduce fresh elements.

* * *

UNTIL the close of the stone age the movements of people had been by means of gradual drifting. During the pal?olithic age, when the population supported itself by hunting, the people wandered over considerable areas in search of game, and the inhabitants of different regions frequently met and mingled with one another. As the forest conditions arose during the close of the Magdalenian period these wanderings were restricted in scope, and with the gradual introduction of domesticated animals and the practice of agriculture during the neolithic age, more settled communities arose.

Thus the different types mixed less with one another, and the communities became more specialised, both in type and culture, as their wanderings diminished.

A new method of intermixture was, however, soon to arise, as the practice of commerce developed. It is possible that even during the pal?olithic age, tribes who lived in a region where flint or other suitable material was abundant, or who had become skilled in the fashioning of some advanced type of implement, sometimes bartered their spare products for other commodities. Such operations, if they did exist, must have been very limited in extent, and confined to bartering between neighbouring tribes.

During the neolithic age this simple principle of exchange continued, though it was probably more frequent, since communities had a narrower range, and some must have been living in regions where suitable raw material was scarce or non-existent. Some well favoured regions also had begun to develop regular commerce. The inhabitants of the island of Santorin, the ancient Melos, had before metal was known organised an export trade in obsidian goods, for they held a monopoly of that excellent volcanic glass in the ?gean region[78]. It seems likely, too, that the people of the Lipari islands traded in the same material with south Italy, Sicily and Malta.[79] Some of the natives of the French department of Indre-et-Loire, finding themselves possessed of great quantities of beautiful honey-coloured flint in the neighbourhood of Le Grand-Pressigny, exported implements, both finished and in the rough, to many distant places in France,[80] and the same is probably true of the dwellers on Pen-maen-mawr, if we may judge from the extensive remains of their industry recently discovered at Graig Llwyd.[81] The industry of Le Grand-Pressigny seems, however, like the obsidian trade in the Mediterranean, to belong to the closing phases of the neolithic age, while the Graig Llwyd factory may well date from the bronze age.

So long as these products of local industry were distributed by land, as in the case of Le Grand-Pressigny and Graig Llwyd, the old method of barter from tribe to tribe was possible and doubtless still continued. But when an island, such as Santorin, was the scene of production, such methods were ineffectual, and a definite organisation for export became necessary. To carry goods from one island to a neighbouring isle or to the mainland requires a ship and a crew, besides some representatives of the makers to effect the sales. When the ship has been equipped it is economical to provide a full cargo, and this would be more than one small community would need or could afford to purchase. This leads to trading voyages of some days' or weeks' duration, when the ship can call at a number of ports to meet the needs of many communities. The inland inhabitants have also to be catered for, and the most serviceable ports became in their turn fresh centres of distribution, and need a dep?t under the supervision of a representative of the makers.

Thus, even before the close of the stone age, we see developing, especially in the Mediterranean region, the beginnings of an organised commerce, involving visits paid by ships and their crews to distant ports and foreign communities, and sometimes leading to the establishment of small foreign trading settlements. With the introduction of metal these features increased rapidly, and, as we shall see, before the bronze age had been in existence for many centuries, an extensive trade had grown up, mainly by sea, but sometimes by land as well, so that bronze became known and used over most parts of Europe which were not too remote from the sea to be affected by sea-borne commerce. Thus a considerable mingling of peoples and cultures took place, not by the sudden arrival of large numbers of invading hordes, but by the constant infiltration of small bodies of merchants and seamen.

The origin of the discovery of metal is still unknown, though many ingenious suggestions have been made. All investigators are agreed that gold, being the most strikingly conspicuous metal, was the first to be noticed and used, though there are those who believe that copper was almost if not quite as early a discovery. Professor Elliot Smith has made interesting suggestions in both cases. He believes that somewhere on the African shore of the Red Sea a cult arose which involved the use of the cowry shell as an amulet for fertility; such cults are well-known and widely spread.[82] For some reason the shells did not ultimately satisfy the people, or the supply diminished, and they made models in gold, deposits of which were found in that locality. Thus the virtue of the amulet, residing originally in its form, became transferred to the material, and gold became and has since remained a lucky and fortunate possession.[83]

Copper, on the other hand, he believes to have been discovered in Egypt. The inhabitants of this country had, in neolithic days, been in the habit of mining malachite in the Sinaitic peninsula, and grinding this mineral on slate palates into a powder, which they applied to their eyes. Green powder thus applied is said to save the wearer from the ill-effects of glaring sunlight, and perhaps served also to keep away the flies, which are a constant source of ophthalmia. Professor Elliot Smith suggests that an Egyptian grinding his lump of malachite on his decorated slate palate, one day met with an unusually hard lump, which he could not grind satisfactorily. In a fit of temper he threw the offending morsel into the fire, doubtless with words of objurgation; later on in the ashes he found a small red bead of copper. A repetition of this action, no doubt with the same formula, produced an identical result, and so the discovery of the reduction of copper from its ore was made.[84] I must admit that at one time I doubted the possibility of this explanation, as I questioned whether the heat of a fire of dung, now and probably then also, the only available fuel, would be sufficient to reduce the ore. To satisfy me on this point, Mr. R. H. Rastall of Cambridge kindly made a laboratory experiment upon a piece of malachite, and as a result assured me that the heat of a dung fire would be ample for the purpose.

While admitting that Professor Elliot Smith's theories are both possible and suggestive, I feel inclined to offer another, albeit more prosaic solution to these problems. Primitive men, whether in prehistoric times or among backward peoples to-day, and, dare I say it, this is perhaps more true of primitive women, have a habit, not, I believe, quite extinct even in more advanced circles, of collecting small objects with natural perforations, or through which holes could readily be drilled, and stringing them upon a thread or wire to make necklaces or bracelets for the adornment of their persons. Such customs carry us back a long way. The old Grimaldi woman from the Grotte des enfants wore two bracelets composed of perforated shells, while her son, if indeed he were her son, had worn on his head a chaplet of the same materials.[85] The Alpine inhabitants of the North Italian lake-dwellings used the vertebr? of pike for the same purpose.[86] Whether the use of strings of beads originated in some religious practice I know not, for it may be that such religious associations, though found to-day among Buddhists, Moslems and Christians, may be relatively modern. That in later days it proved a safe and convenient way of storing accumulated wealth seems more certain, and for this purpose the custom is still practised. Perhaps after all the Preacher was right and this, like everything else, was vanity.

Leaving the cause unsolved, we may be content to note that the practice dates from the first arrival of modern man in Europe, and may be much older. But shells and the vertebr? of fish are easily damaged, and store would have been set by perforated stones, which would have been much more durable. Pebbles of clear quartz, with natural perforations, were worn sometimes by our Saxon forefathers,[87] but such stones are scarce and would have been prized accordingly.

I picture to myself the first discoverer of gold as a young man wishing to obtain the favour of a maid, or perhaps to purchase her from her father. I imagine such a youth going in search of some object, rare and durable and capable of being strung on a necklace. Walking down to a clear stream, perhaps to wash, though more probably to drink or to fish, he noticed in its bed a brilliant yellow stone of quite exceptional beauty. Picking it up and examining it he found he could bend it where it was thin, so that with the aid of a stone he was able to fashion it into the much sought-for bead. Here he had something which was perforated, strong, rare and also beautiful. We can imagine that his success would have been assured. Then would have followed the first gold rush.

Now copper, too, is found in a native state, and is also malleable and easily modelled with a stone hammer; it, too, is capable of exhibiting a bright metallic lustre when clean. Though it could not compare with gold for beauty, or in the permanence of its natural lustre, it could well take second place, and being less rare it soon came to be used freely for decorative purposes. At first it was obtained only in a native state, and was hammered, not melted, as was the case until recent times around Lake Superior.[88] Later some copper ornaments probably fell into the fire, and it was thus discovered that it could be melted. Later still experiments were made with other metallic-looking ores, such as chalcopyrite, and the metal age had come.

Where these discoveries were made is still a matter of uncertainty. Copper objects have been found not uncommonly in tombs of the second predynastic period in Egypt, and sometimes in those of the first.[89] So rare, however, are they in the latter, that, since the two cultures must to some extent have overlapped, it seems possible that the knowledge of this metal was introduced into Egypt by the second pre-dynastic people. It has been suggested recently that these people, with a copper culture, bringing the knowledge of wheat and the cult of Osiris, came from North Syria, from somewhere between Damascus and Beyrut,[90] and if Breasted's views upon the Egyptian calender are sound, we may expect that they entered the Delta 4241 B.C. or thereabouts.[91]

In Mesopotamia we are not very sure of our dates at so early a period, nor have we got any clear evidence of the earliest copper civilisation of that region, but the beautiful copper lions brought back from Tell-el-'Obeid, near Ur, by Dr. H. R. Hall[92], show that at the time when they were made the art of working copper must long have been known, and Dr. Hall tells me that their date may be placed with fair certainty between 3500 and 3000 B.C. Small foundation figures, cast solid in copper, have been found which date from the time of Ur-Nina, about 3000 B.C.[93], while Mr. Raphael Pumpelly, describing his excavations at Anau in Turkestan, states that he found copper implements in a deposit, which on other grounds he dates between 8000 and 7000 B.C.[94] While there is no doubt that copper was found in the lowest layer of the Anau village site, there are few people who agree with the early date claimed for it by Mr. Pumpelly. Taking all the available evidence into consideration, it seems likely that copper was known and used in western Asia as early as 4500 B.C. and might conceivably have been known as early as 5000 B.C.; that it was known before that seems unlikely as far as we can judge from the evidence available at present.

Gold, as we have seen, seems to have been the first metal to be discovered, though we have no sufficient reason for believing that its discovery preceded that of copper by any considerable period. Objects of gold have been found in graves of the second pre-dynastic period in Egypt, as well as some of silver and of lead,[95] so that before 3500 B.C. the metal age had passed its infancy.

It will be seen from what has gone before, that the discovery of metal must have taken place somewhere in western Asia; in Asia Minor, Armenia or Persia. The knowledge spread first of all from tribe to tribe and from city to city, and the objects were traded like the stone axes of Le Grand-Pressigny and Graig Llwyd; but about 4241 B.C. this knowledge was carried into Egypt with an invading people. So far, then, there is no evidence of organised trade, for the gold, silver and lead, which have been found in the pre-dynastic tombs, may have arrived in the same way. Gold, it is true, was found in quartz veins in the granite mountains by the shores of the Red Sea, and in the Wadi Foakhir,[96] but it is not clear that these sources were tapped before the fourth or fifth dynasties; in later days the principal source of supply was Nubia which had, however, been inaccessible to traders until Mernere had made the first cataract passable for navigation about 2570 B.C.[97] Silver was always imported from abroad, probably from Cilicia.[98]

There are reasons for believing that some, at any rate, of the gold used during the period of the Old Kingdom was of foreign origin. Professor Flinders Petrie tells me that Dr. Gladstone made for him an analysis of the gold object found in the tomb of King Khasakhemui, of the second dynasty, who reigned, according to the chronology we are using, about 3200 B.C. He found on this gold object a red crust, which he stated was antimoniate of gold. Now it appears that antimony will only combine with gold in the presence of tellurium, and Professor Petrie tells me that he has been advised that there is no known source of this ore, telluride of gold and antimony, except in Transylvania. I have been informed that all the gold found within the Carpathian ring is of this nature, but as the richest sources lie in Transylvania, where gold was worked by the Romans, the conclusion is the same, that before 3200 B.C. the Egyptians were obtaining gold from Central Europe.

As it seems unlikely that gold would be carried between such distant points as the valleys of the Danube and the Nile by the old method of bartering from tribe to tribe, especially since there are so many physical obstacles on the route, including the Taurus range, it seems more likely that we should see here evidence for an organised sea commerce. Not that I would imply a direct sea traffic from the Danube to the Nile, but that some intermediate people, probably some islanders in the ?gean, the people perhaps of Melos or Crete, traded on the one hand with settlements near the mouth of the Danube and with those in the Delta as well. The obsidian trade of Melos may well be as early as this, in fact it seems to have been on the decline by 3000 B.C., and we find Cretan trade flourishing only a few centuries later. Either or both of these islands might well have been responsible for this traffic.

Oversea trade, then, was in existence, if not very highly developed, during the early days of metal, the centuries preceding 3000 B.C. The knowledge of copper, and the possibility of making copper nails and wire, must have given a great impetus to ship building, which must at this stage have passed from the use of rafts and dug-outs to that of boats built as we know them now. But a new discovery, greater even in some respects than those which I have been describing, was still further to encourage oversea traffic.

The manufacture of implements of flint and obsidian had reached a high pitch of perfection during the early days of metal, and although the new materials were valuable for ornaments, copper knives were, in many respects, less serviceable than stone ones, as the metal is soft and its edge easily turned. It is true that many men, particularly those who wished to display their wealth, preferred copper daggers to those made of flint, for they were more ornate, more novel and had a scarcity value. Those, however, who were poor, or untouched by the fashionable snobbery, preferred the well-tried flint article, which was probably more effective for its purpose.

But with the discovery that the addition of about ten per cent. of tin to the copper produced an alloy of considerable hardness and no little toughness as well, from which could be made implements which seldom chipped or turned, and which could have their edges quickly renewed by hammering or grinding did such an accident happen, the days of copper came quickly to an end, and the traffic in flint implements, even in obsidian, fell upon evil days. It was this discovery, which made metal not merely a luxury, but a really serviceable article to man, which brought the stone age to an end and ushered in the true metal age.

How, when and where this discovery was made is still a mystery. At one time I was disposed to think that it was perhaps in Spain, where both these metals are found, that the discovery was accidentally made, but evidence which has come to hand quite recently has disposed of this idea. Professor Sayce has recently published an extract from a tablet found in the royal library of Assur.[99] It is from a document drawn up in the reign of Sargon of Akkad, whose date has now been finally fixed at 2800 B.C. It is a geographical description of that monarch's empire, giving a list of the provinces, at the close of which it is said that his conquests had extended "from the lands of the setting sun to the lands of the rising sun, namely to the tinland (Ku-Ki) and Kaptara (Crete) countries beyond the Upper Sea (the Mediterranean)." At first there was a tendency to interpret this passage as though Ku-Ki was beyond the Mediterranean, and must refer either to Spain or Brittany; but this is to misunderstand the passage. As Professor Sayce says: "the western extension of the empire ended with the Syrian coast; beyond that were Kaptara or Krete and the Tinland." Ku-Ki may well have been Cyprus, or some other island in the Mediterranean, or some region easily accessible from it.

Now the importance of this passage is that it shows us that as early as 2800 B.C. the Babylonians were cognisant of the existence of tin, and doubtless aware of its value as an ingredient of bronze; this can only mean that they were using it to harden the copper, which they had worked so well centuries earlier. The passage implies that Sargon's rule extended to Ku-Ki, which may perhaps mean no more than that some of his subjects had a trading post there. What seems important is that the discovery of the value of tin and bronze had been made before 2800 B.C., somewhere in western Asia, though at what sites is at present uncertain. Copper mines, which are known to have been worked at an early date, exist south of Trebizonde, near Erzeroum, in Armenia and at Diarbekir in the upper valley of the Tigris; ancient tin workings have been found further east in Khorazan.[100] But the local supply of tin was apparently insufficient, and merchants from the Persian Gulf were carrying on a trade in this commodity with a place in the Mediterranean region, even if they had not already, as seems probable, established a definite trading post in Ku-Ki.

Thus we see that a definite organised trade, both by sea and by land, had been established in the eastern Mediterranean region before 2800 B.C., and that this included a new and important feature, the search for and importation of raw materials as well as the export of manufactured articles.

Now, as I have shown elsewhere,[101] at a date which cannot be very much later, during a period which closed about 2200 B.C., the eastern Mediterranean was in close trade relations with Spain, and was exploiting the mineral resources of that peninsula. At present it is uncertain who these traders were, but they seem to have been in touch with Crete, the Cyclades and the second city of Hissarlik, and perhaps too with Cyprus. Though we have no evidence that these traders were from the Persian Gulf, they were trading between Spain and the area in which Ku-Ki probably lay, and if they were not subjects of the Babylonian Empire, they were at least carrying on the metal trade first organised by the people from the Persian Gulf.

Quite recently it has been stated that there is no clear evidence that the Spanish copper mines had been worked at so early a time,[102] but the data cited by Siret[103] seem to me to prove conclusively that the early settlements of El Argar had direct or indirect trade relations with Hissarlik II., and the discovery throughout the Spanish peninsula of clay balls of a certain type,[104] which exactly resemble some found by Schliemann in the burnt city,[105] seems to me to place this early connection beyond all reasonable doubt.

How early this Spanish trade began we cannot yet say with certainty, beyond the fact that it must have been in existence for some time before the destruction of Hissarlik II. in 2225 B.C.[106] How long it continued in the same hands is also uncertain. But, as I have shown elsewhere,[107] there is evidence that while it lasted, and certainly before 2000 B.C., the eastern traders not only passed through the Pillars of Hercules and discovered the tin fields in the north-west of the peninsula, but learned also that both tin and gold were to be found in the rivers of the south of Brittany. Before the close of the third millennium, probably several centuries before its close, this Levantine trade had reached the Morbihan, where stone axes have been found which repeat the shapes of copper axes from Cyprus.[108]

Now, if we compare the copper and bronze axes found throughout the Mediterranean, from Cyprus to Spain, and those found along the west of Europe from Spain to Brittany, we find a gradual change in form from the triangular axes of Cyprus to the western type, with semi-circular butt and widely splayed edge. The earliest types are found only in the east, the more developed only in the west, for in the east they followed a different line of development. It is true, however, in a general way that the type develops as we pass westward and northward, two or more varieties overlapping at many points en route. This can better be understood by reference to the series of axes shown in Plate I., which could probably be made more perfect, were it possible to get drawings of all the specimens in local museums and private collections.

If again we take the copper daggers, with broad butts and slightly ogival blades, several of which have been found in Crete, and compare them again with those found at Scurgola in South Italy and Monteracello in Sicily,[109] and with other types from Malta,[110] Spain, Brittany and the west, we shall find the type gradually narrowing at the butt and lengthening in the blade, till we come in later centuries to the type commonly known as the rapier, but which I think might more correctly be termed a dirk (see Plate II.).

The gradual evolution of the axe and the dagger as they pass westwards and northwards seems to indicate a line of trade, spreading further and further to the north-west as the centuries pass. At present we must be content with an outline of the movement, but if illustrations of all the specimens found in these regions were available, I doubt not but that the evidence would be more convincing and the details and the dates more minute and exact.

Thus we find these early traders seeking for copper, tin and gold, or any other precious commodities, on the north-west of Europe before 2000 B.C., and it has been shown by various authorities that among the gold-fields explored at that time none was richer than the Irish gold-fields in the Wicklow Hills.[111] It is needless here to recapitulate all the evidence which has been adduced to establish the early working of these deposits. The wealth of gold ornaments of this period found in the island, most of which have passed into the melting pot, but hundreds of which are still in the National Museum at Dublin,[112] would alone be sufficient evidence; but we know also that certain ornaments, known as lunul? or crescents, were exported and reached Brittany, Denmark and Germany.[113] It is likely, too, that gold objects of Irish origin reached to more distant places.[114] This shows us that Ireland was in touch with the trade routes we have been discussing, and this in turn accounts for the vast numbers of bronze implements of early types which are to be found in all museums and private collections, not only in Ireland itself, but throughout Great Britain.

It would seem probable that the early traders from the Mediterranean also reached the Baltic at about the same date, for we find there, too, an early bronze industry, which, while bearing a close resemblance to Central European models, exhibits also western and Mediterranean types.[115] The search for amber probably induced our traders to go to this distant region, for amber, like the precious metals, was much in request in Mediterranean lands, for again it was a substance from which beads could readily be made. It was probably these traders who carried with them the news of the Irish gold-fields, and in due course other traders, starting out from the Baltic, joined the gold rush. We have already seen that a gold crescent of Irish work has been found in Denmark, we can find, too, other evidence of this trade.

Now if we plot out on a map of the British Isles the sites at which have been found the bronze implements of this period, and such a map of flat celts was published some years ago by Mr. O. G. S. Crawford,[116] we shall notice certain striking features. Where the chalk lands or limestone hills exist these finds are fairly numerous and generally distributed, for, as Crawford showed, these areas were open grass lands. But throughout the rest of the country these sites string out into long lines, and these lines, if produced, would intersect near Dublin; these lines seem to indicate trade routes, passing through thickly wooded and probably uninhabited country on their way to the Irish gold-fields.

One such route starts from Southampton and passing Winchester, crosses the Kennet at Newbury, where it was met perhaps by a route from Chichester. Thence it passed by the head waters of the Thames to a point on the Cotswolds not far from Cirencester, where it may have been joined by other routes from the south-west. It descended the scarp slope of the Cotswold at or near Broadway, crossed the Avon near Evesham, and the Severn at Bevere Island above Worcester. Thence it passed up the west side of the valley, crossing the river again below Shrewsbury. Its course across north Shropshire seems to have lain on the watershed between the Tern and the Perry, if we may judge from evidence of a later date,[117] thence passing from Ebnal towards Llanarmon-dyffryn-Ceiriog, it crossed over to the Dee Valley, where we can pick up fresh evidence near Corwen. From the head of Bala lake it seems to have turned slightly north of west, instead of passing down the Mawddach Valley, and it reached the coast somewhere to the north of Harlech, perhaps by the so-called Roman steps at Cwm Bychan. This is the best attested route so far traced out, but further work is required to establish its course with precision all the way.

Another route from the Yorkshire coast through York to the Aire gap has been described by Colonel E. Kitson Clark,[118] while some years ago I traced several from the borders of the Fens into Leicestershire, where they met at Bardon Hill; thence the route passed through Ashby-de-la-Zouch as far as Burton-on-Trent, where it seemed to be pointing to the Peak district.[119] There appears to be a route running thence by Macclesfield and Knutsford towards Warrington, while there are signs that the route through the Aire gap also turned south towards the same spot. Near Warrington a number of flat axes have been found,[120] some on the north and some to the south. The northern settlement was in the parish of Winwick, and among the things found there and dating from this time is a battle-axe of the so-called boat-axe or batyx type.[121] This type and the flint of which it is made both indicate Denmark as the place of origin. The fact that both these trade routes run to Warrington, which seems then to have been an island in the middle of the Mersey, shows, I think, that here we have a port, from which in the early bronze age Baltic traders set sail for Dublin Bay.[122] Warrington, therefore, rather than Chester, was the first predecessor of Liverpool, and the Mersey holds its own as the earliest estuary used for the western trade. Crawford's map also shows that a similar trade route must have crossed Scotland from the Firth of Forth to the Firth of Clyde and the Mull of Galloway, but no details of such route have been worked out.

Much work yet remains to be done before the courses of these trade routes can be traced with precision and their dates fully established, but enough has been said, I trust, to show that in addition to direct sea routes from Brittany, the Irish gold fields tempted traders to cross both England and Scotland on their way from France and the Baltic.[123] These traders would have needed provisions for the journey, and for these would have bartered bronze axes to the people settled on the chalk downs and limestone hills. The journey across the Midland plain was through a densely wooded and probably uninhabited area, and in passing through Wales they kept mostly to the valleys, while the bulk of the population grazed its sheep on the high moorlands.[124] The few axes found must have been such as were lost by the way, and considering the number found this indicates an extensive traffic.

In Ireland the traders probably employed the natives to wash the alluvial gold; they had also to barter with them for their supplies. No wonder, then, that bronze implements of the earliest type have been found almost more abundantly in that island than in any other part of Europe, while the number of gold objects found there is unsurpassed elsewhere. Doubtless the natives worked the gold fields sometimes on their own account, and they seem also to have tried to supply themselves with home-made metal axes. There are veins of copper ore in various parts of the island, which they seem to have discovered, but tin is to all intents and purposes absent.[125] It is possible, too, that the traders refused to divulge the secret of the tin alloy. It would have been strange indeed had they not done so, and so the native Irish, for a time at least, made themselves axes of copper. This, at least, seems so be the most plausible explanation of the great number of copper axes found in that island.

The foregoing is, of necessity, but a brief account of the early metal trade and its relations with Celtic lands. To do the subject justice would require more space than is at my disposal; nor is the time yet ripe for more detailed treatment. This outline will serve to show that foreign elements reached Celtic lands some 4000 years ago, though in small numbers; who these people were must be considered in the next chapter.

* * *